Shows us why visualizing residuals from a model is important

orly-owl-residuals

library(broom)
library(ggplot2)

# read in the data
# others available here: http://www4.stat.ncsu.edu/~stefanski/NSF_Supported/Hidden_Images/stat_res_plots.html
owl <- read.table("http://www4.stat.ncsu.edu/~stefanski/NSF_Supported/Hidden_Images/orly_owl_files/orly_owl_Lin_4p_5_flat.txt", 
    header = FALSE)

# fit the linear model
fit <- lm(V1 ~ . - 1, data = owl)

# save the residuals via broom
owl_vars <- augment(fit)

# plot fitted vs residuals
ggplot(owl_vars, aes(x = .fitted, y = .resid)) + 
    geom_point(size = 1)